Documente Academic
Documente Profesional
Documente Cultură
com/
Examining the Reliability and Validity of Clinician Ratings on the Five-Factor Model Score Sheet
Lauren R. Few, Joshua D. Miller, Jennifer Q. Morse, Kirsten E. Yaggi, Sarah K. Reynolds and Paul A. Pilkonis Assessment 2010 17: 440 originally published online 2 June 2010 DOI: 10.1177/1073191110372210 The online version of this article can be found at: http://asm.sagepub.com/content/17/4/440
Published by:
http://www.sagepublications.com
Additional services and information for Assessment can be found at: Email Alerts: http://asm.sagepub.com/cgi/alerts Subscriptions: http://asm.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://asm.sagepub.com/content/17/4/440.refs.html
>> Version of Record - Oct 27, 2010 Proof - Jun 2, 2010 What is This?
Examining the Reliability and Validity of Clinician Ratings on the Five-Factor Model Score Sheet
Lauren R. Few1, Joshua D. Miller1, Jennifer Q. Morse2, Kirsten E. Yaggi2, Sarah K. Reynolds2, and Paul A. Pilkonis2
Assessment 17(4) 440453 The Author(s) 2010 Reprints and permission: http://www. sagepub.com/journalsPermissions.nav DOI: 10.1177/1073191110372210 http://asmnt.sagepub.com
Abstract Despite substantial research use, measures of the five-factor model (FFM) are infrequently used in clinical settings due, in part, to issues related to administration time and a reluctance to use self-report instruments. The current study examines the reliability and validity of the Five-Factor Model Score Sheet (FFMSS), which is a 30-item clinician rating form designed to assess the five domains and 30 facets of one conceptualization of the FFM. Studied in a sample of 130 outpatients, clinical raters demonstrated reasonably good interrater reliability across personality profiles and the domains manifested good internal consistency with the exception of Neuroticism. The FFMSS ratings also evinced expected relations with self-reported personality traits (e.g., FFMSS Extraversion and Schedule for Nonadaptive and Adaptive Personality Positive Temperament) and consensus-rated personality disorder symptoms (e.g., FFMSS Agreeableness and Narcissistic Personality Disorder). Finally, on average, the FFMSS domains were able to account for approximately 50% of the variance in domains of functioning (e.g., occupational, parental) and were even able to account for variance after controlling for Axis I and Axis II pathology. Given these findings, it is believed that the FFMSS holds promise for clinical use. Keywords DSM-5, personality disorders, assessment One of the most widely researched and accepted models of personality is the five-factor model (FFM; Digman, 1990), which includes the following five broad dimensions: Neuroticism (vs. Emotional Stability), Extraversion (vs. Introversion), Openness (vs. Closedness to Experience), Agreeableness (vs. Antagonism), and Conscientiousness (vs. Disinhibition). In addition to these five broad domains, Costa and McCrae (1992) delineated six underlying facets subsumed by each of the FFM domains and assessed in their popular measure of the FFM, the Revised NEO Personality Inventory (NEO PI-R; Costa & McCrae, 1992). Although the FFM is a well-established model of personality that has proven helpful in integrating findings from various personality models, conceptualizing personality disorder (PD), and bridging the divide between normal and abnormal personality, there is a smaller body of research addressing the clinical utility of this model. The limited research that is available, however, suggests that it has promise as it has been demonstrated that measures of the FFM predict treatment satisfaction and compliance (Miller, Pilkonis, & Mulvey, 2006), as well as functional impairment (e.g., Hopwood et al., 2009; Miller, Pilkonis, & Clifton, 2005). In addition, numerous studies have documented substantial relations between the FFM and Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association, 1994) PDs (see Costa &Widiger, 2002). Finally, clinicians view it as a useful model for conceptualizing personality pathology (Samuel & Widiger, 2006). Despite its potential utility, clinicians have been slow to adopt measures of the FFM in their clinical work. There may be several reasons for this reluctance. First, the most popular measure of the FFM, the NEO PI-R (Costa & McCrae, 1992), is a 240-item measure that takes approximately 20 to 40 minutes to complete.1 The NEO PI-R can be completed using self- or informant ratings or interview (Structured Interview for the Five-Factor Model [SIFFM]; Trull & Widiger, 1997), but the self-report methodology is
1 2
University of Georgia, Athens, GA, USA Western Psychiatric Institute and Clinic, University of Pittsburgh Medical Center, Pittsburgh, PA, USA Corresponding Author: Joshua D. Miller, Department of Psychology, University of Georgia, Athens, GA 30602-3013, USA Email: jdmiller@uga.edu
Few et al. most common. Although both self and informant ratings of personality demonstrate predictive validity (Klein, 2003), informant reports often provide stronger predictive or incremental validity over self-report data (Miller et al., 2005). Another issue that may limit the perceived clinical utility of FFM assessments is that these measures do not assess for impairment in relation to the various traits. This omission is significant as it has been argued that an explicit assessment of impairment is necessary when assessing PDs (Clark, 2007). Thus, despite the evidence that personality data derived from an FFM assessment show promise with regard to patient conceptualization, prediction of treatment outcomes, and level of impairment, the current assessment methodology may be regarded as too unwieldy and limited to warrant frequent use in clinical settings. Given these concerns, several alternative assessment strategies have been developed that address some of these problems. For instance, the SIFFM (Trull & Widiger, 1997) is an interview-based method for assessing the traits consistent with the NEO PI-R that also includes an assessment of the dysfunction associated with each trait, which may increase its clinical utility. Unfortunately, the length of this interview may have limited its adoption in clinical settings. A recently developed measure, the FFM Rating Form (FFMRF; Widiger, 2004), has addressed concerns regarding brevity and the need to include lower order facets in other-rated assessment of the FFM. The FFMRF uses a single-item to assess each of the 30 facets of the FFM by including an identifying term for each facet (e.g., selfconsciousness), along with two to four adjectives that describe both poles of each facet (e.g., timid, embarrassed vs. self-assured, glib, shameless). The FFMRF has been reliably used to describe prototypical cases of PD in terms of the FFM (Lynam & Widiger, 2001; Samuel & Widiger, 2004), and self-reports on the FFMRF demonstrated good convergent validity with self-reports on the NEO PI-R among undergraduate samples (Mullins-Sweatt, Jamerson, Samuel, Olson, & Widiger, 2006). The FFMRF has also been used to compare the clinical utility of the FFM with the DSM-IV model (Lowe & Widiger, 2009). Another brief measure of the FFM is the FFM Score Sheet (FFMSS; Widiger & Spitzer, 2002), which is an earlier version of the FFMRF that also uses a single rating for each of the 30 personality facets captured in the NEO PI-R.2 Although the content of both the FFMSS and FFMRF is largely the same, the two differ in that the FFMSS uses a scale ranging from 1 (problematic, very low on the trait) to 7 (problematic, very high on the trait; see the appendix), whereas the FFMRF uses a 5-point scale ranging from extremely low to extremely high. The FFMSS also allows for ratings that convey high or low levels of the trait that are not problematic (e.g., 3, low on the trait; 5, high on the trait). This difference is noteworthy because the FFMSS scale
441 includes indications of poor adaptations at the poles of each FFM facet; this difference may be important in that previous research suggests that the NEO PI-R does not contain much content related to the maladaptive poles of certain dimensions (e.g., maladaptively high Agreeableness; Haigler & Widiger, 2001). Haigler and Widiger demonstrated that rather straightforward alterations of some of the NEO PI-R content to increase the description of maladaptivity at the high end of certain dimensions (i.e., Openness, Agreeableness, Conscientiousness) resulted in stronger correlations among certain domains (i.e., Conscientiousness) and PDs (i.e., ObsessiveCompulsive PD). The goals of the current study were multifold. First, we tested the interrater reliability and internal consistency (domains only) of the ratings derived from the FFMSS. Second, we examined which FFM traits were more/less difficult for clinicians to rate reliably. Third, we examined the convergent and discriminant validity of the FFMSS ratings in relation to self-report personality traits and expert ratings of the DSM-IV PDs. Fourth, we examined the utility of the FFMSS ratings by examining their relations with expert ratings of functional impairment (e.g., interpersonal, occupational). We also tested whether the FFMSS ratings provided incremental validity in the statistical prediction of these impairment domains above and beyond that provided by ratings of depression, anxiety, and DSM-IV PD symptoms. Also unique to the current study is that it provided the first test of these single-item facet ratings for the assessment of the FFM in a clinical setting. In previous research with the FFMRF, the measure has been used to generate ratings of prototypical PDs (Samuel & Widiger, 2004), PD vignettes (Sprock, 2002), or self-report descriptions of undergraduates (Mullins-Sweatt et al., 2006). We expand on each of these goals below. As noted above, in the current study, we investigated which FFM traits were most difficult to rate using the FFMSS. Research examining self- and other ratings of personality may be valuable in evaluating the potential difficulty for clinicians in rating FFM dimensions using the FFMSS. For example, in a clinical sample of predominantly depressed individuals, Bagby et al. (1998) found that Openness and Neuroticism were rated less reliably relative to the other FFM domains, and Conscientiousness and Agreeableness were rated most reliably, followed by Extraversion. A similar pattern of findings is hypothesized here. In the current study, we also tested the convergent and discriminant validity of the FFMSS in relation to personality as measured by the Schedule for Nonadaptive and Adaptive Personality (SNAP; Clark, 1993) trait and temperament scales and expert rating of DSM-IV PD symptoms. The relations between the FFMSS and the SNAP were expected to be largely consistent with those found by Clark, Vorhies, and McEwen (2002). For example, FFMSS
442 Extraversion should correlate significantly with SNAP Positive Temperament, Exhibitionism, Detachment, Entitlement, and Impulsivity, whereas FFMSS Neuroticism should be significantly related to SNAP Negative Temperament, Self-Harm, Mistrust, Dependency, and Aggression. FFMSS Conscientiousness should be significantly related to Disinhibition, Workaholism, Impulsivity, Manipulativeness, Dependency, and Propriety. The additional FFM factors with no explicit counterpartAgreeableness and Opennessshould also demonstrate significant correlations with certain SNAP scales. For instance, FFMSS Agreeableness should be related to SNAP Aggression, Mistrust, Manipulativeness, and Detachment. Finally, FFM Openness should be positively related to SNAP Eccentric Perceptions and Impulsivity and negatively related to Propriety. With regard to the relations between the FFMSS and DSM-IV PDs, we expected that the findings should be consistent with theoretical conceptualizations of PDs from a FFM perspective (Lynam & Widiger, 2001). For instance, individuals with greater Borderline PD symptoms should be rated higher on the FFMSS Neuroticism facets of anxiety, angerhostility, depression, impulsiveness, and vulnerability, and they should be rated lower on FFMSS Agreeableness facets of trust and compliance, in addition to the competence facet of Conscientiousness. To quantify these relations, we examined the similarity of the FFMSSDSM-IV PD correlates found in the current sample with the hypothesized correlates between the FFM and DSM-IV PDs generated by expert raters (Lynam & Widiger, 2001). Finally, we examined the association between FFMSS ratings and impairment (i.e., romantic relationships, parenting, other social relationships, occupational impairment, distress caused to significant others, and overall impairment). Consistent with previous results, we expected that the impairment domains would be positively related to Neuroticism and negatively related to Extraversion, Agreeableness, and Conscientiousness (e.g., Hopwood et al., 2009; Miller et al., 2005). The results from these analyses are important in testing the clinical utility of the FFMSS ratings; that is, do clinicians ratings of these personality traits correspond to functional impairment in a variety of different domains? We also tested the incremental validity of the FFMSS by evaluating whether it accounted for additional variance in the prediction of impairment beyond Axis I symptoms (i.e., depression, anxiety) and Axis II PD symptoms. This analysis provided important information regarding whether the FFMSS ratings provide unique information that is not captured by the current DSM constructs as has been demonstrated previously with FFM information derived from other sources (i.e., Miller et al., 2006).
Assessment 17(4)
Few et al. The Longitudinal Expert Evaluation Using All Data (LEAD) method (Spitzer, 1983) was used in determining all consensus ratings. This method emphasizes the contribution of expert clinical judgment, but includes the use of multiple information sources in arriving at that judgment. These sources included assessment interviews with the patient as well as judgments of other professionals. Following the assessment sessions, the primary interviewer presented the case at a 3- to 4-hour diagnostic conference with colleagues from the research team. A minimum of three judges participated.3 All available data were reviewed and discussed at the conference. Judges were given access to all data that had been collected: current and lifetime Axis I information, symptomatic status, social and developmental history, and personality features acknowledged on the Axis II interviews. The self-report personality data collected with the SNAP were not available for the LEAD ratings. The relevant data for the current study that were derived from the case conference include (a) consensus ratings of DSM-IV PD criteria and (b) consensus ratings on impairment variables. The FFMSS ratings were typically completed by the primary interviewer prior to the case conference, whereas the secondary FFMSS rater (where available) did so following the completion of the case conference (the FFMSS ratings were not discussed during the case conference). All FFMSS raters received basic training in relation to the FFM and orientation to the FFMSS, which involved reviewing the FFM facets using descriptions from Costa and McCraes (1992) NEO PI-R manual. Raters also received copies of these descriptions that could be used as a reference when completing the FFMSS.
443 ranged from .71 (Disinhibition) to .91 (Aggression). The SNAP traits have shown to be related to DSM-IV PDs in theoretically expected directions (e.g., Reynolds & Clark, 2001). Consensus ratings of DSM-IV PD criteria. These ratings were determined in each participants case conference. A consensus rating of each DSM-IV PD symptom was determined using a 0-2 scale, with 0 indicating absent, 1 indicating present, and 2 indicating strongly present. Symptom counts for each participant were generated by adding all scores (i.e., 0, 1, 2) for each PD. Alpha coefficients for the PDs ranged from .53 (Dependent PD) to .88 (Avoidant PD), with a median of .73. Ratings of depression and anxiety. These ratings were conducted with the Hamilton Depression Rating Scale (HAM-D; Hamilton, 1960) and the Hamilton Anxiety Rating Scale (HAM-A; Hamilton, 1959) by the primary interviewer for each participant. Studies have supported the reliability of HAM-D ratings by clinicians (Carroll, Fielding, & Blashki, 1973). The HAM-A has also been shown to be both reliable and valid when used with individuals with depressive disorders and anxiety disorders (Maier, Buller, Philipp, & Heuser, 1988). Consensus ratings of impairment. Consensus ratings were determined separately for romantic relationships, parenting, other social relationships (e.g., friends, family members), occupational impairment, distress caused to significant others (e.g., friends, children), and overall impairment using a 1 (exceptionally positive functioning) to 9 (difficulties are persistent and pervasive, without clearly identifiable elements of functioning relevant to the domain) one-item scale with higher scores indicative of greater impairment.
Measures
Five-Factor Model Score Sheet. The FFMSS (Widiger & Spitzer, 2002; see the appendix) is a one-page rating sheet consisting of 30 items representing each of the 30 facets of the FFM, as conceptualized in the NEO PI-R. These facets are organized with respect to the FFM domains, such that there are six items beneath a listed domain. Each item includes a list of two to four adjectives describing the trait. For example, high straightforwardness, a facet of Agreeableness, is described with the adjective naive whereas low straightforwardness is described using the adjective deceptive. Each item is rated on a 1 (problematic, very low on the trait) to 7 (problematic, very high on the trait) scale. Of the 130 participants, 112 were rated by two raters. Schedule for Nonadaptive and Adaptive Personality. The SNAP (Clark, 1993) is a 375-item, truefalse inventory that assesses 15 traits relevant to PD: 12 lower order primary traits and three broad temperament dimensionsNegative Temperament, Positive Temperament, and Disinhibition (vs. Constraint). Alpha coefficients in the current study
Results
Prior to completing reliability and validity analyses, descriptive statistics for the FFMSS were examined. Means and standard deviations for the facets and domains are presented in Table 1. Intercorrelations between FFMSS domains were also computed. Several analyses were conducted to examine the reliability of the FFMSS. First, internal consistency of the FFMSS domains was examined. Second, individual profile agreement, or reliability of ratings for each participant on the FFMSS facets, was determined. Third, interrater reliability of ratings across the FFMSS domains and facets was examined. Following the reliability analyses, several analyses were conducted to examine convergent and divergent validity. The FFMSS ratings (using only the primary raters scores) were examined in relation to (a) SNAP scales, (b) consensus DSM-IV PD ratings, and (c) consensus impairment ratings across a variety of domains (e.g., romance, work). To control for Type I error, we lowered our significance level to p .001 for all analyses.
444
Table 1. Descriptive Statistics and Interrater Reliability Coefficients for FFMSS Traits M Neuroticism (a = .61) Anxiety Angry hostility Depression Self-consciousness Impulsiveness Vulnerability Extraversion (a = .88) Warmth Gregariousness Assertiveness Activity Excitement seeking Positive emotions Openness (a = .87) Fantasy Aesthetics Feelings Actions Ideas Values Agreeableness (a = .86) Trust Straightforwardness Altruism Compliance Modesty Tender-mindedness Conscientiousness (a = .92) Competence Order Dutifulness Achievement-striving Self-discipline Deliberation 30.85 5.52 4.88 5.35 5.06 4.69 5.35 22.82 3.89 3.71 4.08 3.71 4.00 3.42 24.89 4.12 4.24 4.65 3.94 3.89 4.06 23.85 3.60 4.25 4.07 3.78 4.09 4.06 23.07 3.90 3.91 4.02 3.58 3.64 4.02 SD 4.22 1.23 1.27 1.12 1.46 1.38 1.10 6.83 1.44 1.62 1.40 1.39 1.41 1.34 6.35 1.49 1.25 1.41 1.49 1.31 1.24 5.98 1.52 1.22 1.27 1.34 1.30 1.08 6.52 1.34 1.26 1.34 1.20 1.26 1.34 ICCDE .55 .48 .54 .30 .59 .54 .56 .68 .50 .62 .56 .62 .62 .61 .58 .51 .38 .41 .47 .45 .50 .66 .53 .49 .47 .59 .60 .49 .72 .57 .45 .52 .37 .54 .51 Conv r .43* .30* .45* .38* .06 .54* .54* .83* .60* .78* .64* .78* .64* .73* .59* .67* .72* .64* .55* .74* .71* .71* .55* .71* .79* .75* .82* .75* .79* .76*
Assessment 17(4)
Median Div r .12 -.20 -.13 .01 .03 -.07 .18 .04 -.11 .09 -.14 .08 .00 .06 .11 -.06 .01 -.01 .12 .06 .10 .04 .04 .16 .01 -.01 .02 .08 .01 -.09
Note: FFMSS = Five-Factor Model Score Sheet; ICCDE = double entry intraclass correlation coefficient; N = 130 for descriptive statistics; N = 112 for ICCDE; Conv r = corrected itemtotal correlations between a specific facet and the home domain score without that facet (e.g., Compliance Conv r = correlation between Compliance and Agreeableness without the Compliance facet included in the total domain score); Median Div r = median correlation of each facet with facets from the other four FFM (five-factor model) domains. *p .001.
To examine individual profile agreement, a double-entry intraclass correlation coefficient (ICCDE) was used as it has proven to be a superior measure for calculating profile agreement (see McCrae, 2008 for a review). The ICCDE was computed for 112 of the 130 participants (all those who had data from two raters). One benefit of this approach is that it takes into consideration absolute agreement of the ratings (rather than just the relative agreement) generated by each rater. To do this, each of the raters 30 facet ratings for a given participant was entered. Then those same ratings were pasted into the other raters column (e.g., Rater 1 has his or her 30 FFMSS ratings of a participant listed followed by Rater 2s 30 FFMSS ratings of that same participant in
Few et al.
Table 2. Interrelations Between the Five-Factor Model Score Sheet Domains Neuroticism Neuroticism Extraversion Openness Agreeableness Conscientiousness
*p .001.
445
Openness
Agreeableness
Conscientiousness
.02 -.18
.41*
Table 3. Correlations Between Clinician-Rated FFMSS Traits and SNAP Scales Neuroticism Negative temperament Mistrust Manipulativeness Aggression Self-harm Eccentric perceptions Dependency Positive temperament Exhibitionism Entitlement Detachment Disinhibition Impulsivity Propriety Workaholism .45* .23 .02 .13 .51* .10 .27* -.47* -.29 -.23 .36* .11 .10 .05 .01 Extraversion -.11 .04 .03 .16 -.06 .19 -.05 .52* .42* .18 -.47* .12 .14 .08 .07 Openness .10 .02 .20 .13 .21 .32* .05 .20 .25 .08 -.07 .17 .30* -.07 .03 Agreeableness -.06 -.22 -.27 -.38* -.11 -.16 .08 .05 -.03 -.25 -.13 -.24 -.15 .06 .06 Conscientiousness -.30* -.28 -.34* -.33* -.31* -.31* -.15 .09 .03 -.16 -.09 -.32* -.28 .05 .30*
Note: FFMSS = Five-Factor Model Score Sheet; SNAP = Schedule for Nonadaptive and Adaptive Personality. N = 125. *p .001.
column 1; Rater 2 has his or her 30 FFMSS ratings of that participant listed followed by Rater 1s 30 FFMSS ratings of that same participant in column 2). Following this data manipulation, a Pearson correlation was computed between these columns (with 60 data points in each column). For the 112 participants, the interrater reliability coefficients for the entire individual FFM profiles ranged from -.31 to .92 with a median coefficient of .58. We also examined which traits were more difficult to rate reliably using the same double-entry method (Table 1). Coefficients for the facets ranged from .30 (depression) to .62 (gregariousness, excitement-seeking, activity) with a median coefficient of .52. Coefficients for the domains ranged from .55 (Neuroticism) to .72 (Conscientiousness) with a median of .66.
NEO PI-R, the intercorrelations in the current study were compared with the intercorrelations of the NEO PI-R domains reported in the NEO PI-R normative sample (Costa & McCrae, 1992). To do this, the two sets of correlations were tested to see if there were any significant differences (i.e., test for independent rs; Cohen and Cohen, 1983). Of the 10 pairs of correlations (e.g., FFMSS Neuroticism and FFMSS Extraversion vs. NEO PI-R Neuroticism and NEO PI-R Extraversion), two were significantly different (z 3.44, p .001): Conscientiousness and Neuroticism (NEO PI-R, -.53; FFMSS, -.36) and Conscientiousness and Extraversion (NEO PI-R: .27; FFMSS: -.05) such that the FFMSS domains appear to be more orthogonal than the NEO PI-R domains.
446 counterparts (Negative Temperament, r = .45; Positive Temperament, r = .52; Disinhibition, r = -.32). With regard to all the SNAP scales, FFMSS Neuroticism was significantly positively related to Self-Harm, Dependency, and Detachment and negatively related to SNAP Positive Temperament. FFMSS Extraversion was significantly positively related to SNAP Exhibitionism and negatively related to Detachment. Significant positive correlations emerged between FFMSS Openness and SNAP Eccentric Perceptions and Impulsivity, whereas significant negative correlations emerged between FFMSS Agreeableness and SNAP Aggression. Last, FFMSS Conscientiousness was positively related to SNAP Workaholism and negatively related to SNAP Negative Temperament, Manipulativeness, Aggression, Self-harm, and Eccentric Perceptions. Overall, FFMSS Neuroticism and Extraversion demonstrated the largest correlations with Positive Temperament (-.47 and .52, respectively). FFMSS Openness manifested its largest correlation with Eccentric Perceptions (.32), whereas FFMSS Agreeableness and Conscientiousness evinced the largest correlations with Aggression (-.38) and Manipulativeness (-.34), respectively.
Assessment 17(4) (OCPD) with a median of .85 (Table 4, final row). Overall, the FFMSSDSM-IV PD correlates were quite similar to the expert descriptions of prototypical individuals with each PD.
Discussion
The goal of the current study was to evaluate the reliability and validity of the FFMSS (Widiger & Spitzer, 2002). Previous research has identified a number of reliable and valid measures of the FFM (e.g., NEO PI-R, NEO-FFI, Costa & McCrae, 1992; FFMRF, Widiger, 2004). There are various aspects of these measures, however, which may contribute to their limited use in clinical settings, such as administration time (e.g., NEO PI-R; SIFFM), self-report methodology (e.g., NEO-FFI; NEO PI-R), and limited breadth (i.e., assess domains only: NEO-FFI). Given the likelihood that the DSM-V may include a dimensional trait model of personality for use in the diagnosis and description of personality
Few et al.
Table 4. Correlations Between Clinician-Rated FFMSS Traits and Consensus DSM-IV Personality Disorder Ratings PPD Neuroticism Anxious Angry hostility Depressiveness Self-consciousness Impulsivity Vulnerability Extraversion Warmth Gregariousness Assertiveness Activity Excitement seeking Positive emotions Openness Fantasy Aesthetic Feelings Actions Ideas Values Agreeableness Trust Straightforwardness Altruism Compliance Modesty Tender-mindedness Conscientiousness Competence Order Dutifulness Achievement Self-discipline Deliberation r With L&W .11 -.20 .39 .07 -.16 .20 .11 -.12 -.32 -.13 .22 -.20 .02 -.16 -.07 .04 -.13 -.12 .04 -.05 -.14 -.49 -.44 -.35 -.34 -.39 -.29 -.46 -.27 -.09 -.12 -.28 -.29 -.34 -.26 .47 SPD .13 .22 .03 .21 .18 -.21 .07 -.31 -.32 -.27 -.16 -.19 -.28 -.26 -.20 -.11 -.11 -.19 -.25 -.07 -.16 -.05 -.05 -.03 .01 -.02 -.05 -.10 .10 .03 .13 .03 -.05 .15 .21 .77* STPD .14 .20 .15 .00 -.03 .08 .11 -.12 -.23 -.17 .11 -.01 -.15 -.08 .21 .24 .14 .03 .17 .38 .02 -.22 -.17 -.16 -.28 -.18 -.12 -.12 .02 .13 .09 -.01 .02 -.08 -.05 .53 ASPD .05 -.22 .27 -.15 -.20 .45 -.01 .17 -.09 .07 .26 .04 .42 .13 .20 .11 .04 .11 .31 .12 .24 -.41 -.30 -.43 -.31 -.39 -.20 -.30 -.39 -.28 -.23 -.38 -.32 -.37 -.40 .90* BPD .42 -.03 .44 .07 -.04 .60 .46 .18 .01 .05 .28 .04 .45 .02 .36 .20 .21 .26 .41 .30 .30 -.33 -.20 -.39 -.32 -.34 -.14 -.17 -.45 -.30 -.25 -.41 -.39 -.46 -.48 .84* HPD .09 -.15 .18 -.06 -.17 .29 .27 .36 .19 .37 .29 .21 .37 .28 .31 .28 .15 .31 .30 .21 .17 -.15 .05 -.12 -.18 -.17 -.26 -.05 -.29 -.14 -.15 -.27 -.19 -.30 -.41 .93* NPD -.03 -.13 .29 .02 -.26 .08 -.07 .07 -.24 .03 .29 .13 .07 .06 .21 .43 .25 -.08 .11 .21 .05 -.49 -.17 -.35 -.48 -.33 -.64 -.36 -.07 .05 .03 -.15 .00 -.09 -.21 .86* AVPD .35 .40 -.06 .44 .63 -.31 .15 -.66 -.39 -.62 -.64 -.50 -.53 -.47 -.36 -.31 -.20 -.30 -.44 -.11 -.27 .22 -.10 .20 .14 .23 .45 .16 .18 .12 .20 .13 -.03 .11 .36 .87* DPD .37 .15 .11 .15 .20 .24 .46 -.01 .15 .04 -.16 -.07 .03 -.03 -.02 -.09 -.12 .16 .07 -.03 -.09 .12 .23 -.03 -.04 .03 .18 .14 -.21 -.23 -.13 -.21 -.25 -.13 -.13 .52
447
OCPD .02 .20 .09 .14 .07 -.30 -.10 -.15 -.18 -.16 -.02 -.01 -.27 -.08 -.24 -.09 -.03 -.24 -.30 -.18 -.28 -.08 -.05 -.02 -.02 -.07 -.12 -.09 .39 .42 .32 .28 .36 .28 .33 .95*
Note: FFMSS = Five-Factor Model Score Sheet; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; PPD = paranoid personality disorder; SPD = schizoid personality disorder, STPD = schizotypal personality disorder; ASPD = antisocial personality disorder; BPD = borderline personality disorder; HPD = histrionic personality disorder; NPD = narcissistic personality disorder; AVPD = avoidant personality disorder; DPD = dependent personality disorder; OCPD = obsessivecompulsive personality disorder; r with L&W = r with Lynam and Widiger (2001) expert FFM (five-factor model) PD prototype ratings. Correlations |.28| are significant at the .001 level. N = 130.
Table 5. Correlations Between Clinician-Rated Five-Factor Model Score Sheet Traits and Impairment Overall Neuroticism Extraversion Openness Agreeableness Conscientiousness R2 domains .57* -.28* .06 -.45* -.51* .56* Romantic .39* -.14 .12 -.26 -.31* .23* Parental .26 .10 .13 -.45* -.49* .37* Occupational .45* -.21 .09 -.41* -.62* .52* Social .54* -.44* -.09 -.37* -.39* .50* Distress to Others .52* .02 .16 -.52* -.56* .60*
Note: N = 59 for parental impairment; N = 130 for all other impairment variables; Overall = overall impairment; Romantic = romantic impairment; Parental = parental impairment; Occup = occupational impairment; Social = social impairment. *p .001.
448 pathology, it will be important to develop a clinically reliable, valid, and useful (e.g., efficient) measure of traits such as those found in the FFM. The current study extended research in the assessment of the FFM by evaluating a clinician-rated, single-item rating form of the 30 FFM facets.
Assessment 17(4) Interrater reliability of the individual FFMSS facets and domains was also examined to determine which traits were more reliably rated by clinicians. Based on previous research examining selfother convergence in FFM ratings (e.g., Bagby et al., 1998), it was hypothesized that Openness and Neuroticism would be the most difficult domains to rate reliably, whereas Conscientiousness, Agreeableness, and Extraversion would evince the greatest interrater reliability. Results supported these hypotheses, in that Neuroticism and Openness manifested the lowest interrater reliability, whereas Conscientiousness, Agreeableness, and Extraversion were the domains rated most reliably. These findings are also consistent with previous research examining the interrater reliability of thin-slice ratings of personality (Oltmanns, Friedman, Fiedler, & Turkheimer, 2004). Specifically, when rating military recruits, strangers rated Neuroticism least reliably and Extraversion most reliably. At the facet level, the median ICC was .52. Despite fair to good reliability across the FFMSS domains and 27 of 30 facets, interrater reliability for the FFMSS in the current study was lower than the SIFFM (Trull et al., 1998), an interview-based measure of the FFM. This is not surprising as the SIFFM was designed to assess these traits directly, whereas the FFMSS was completed without direct assessment of these traits. The three facets that exhibited the poorest inter-rater reliability were depression (.30), aesthetics (.38), and achievement-striving (.37), which manifested similarly low correlations between two peer raters (Costa & McCrae, 1992). It is possible that depression was rated less reliably because one rater interviewed the patient, whereas the other made ratings on the basis of case conference information. Given the nature of depression, face-to-face contact with the patient may have influenced the ratings (e.g., assessment of flat or depressed affect; crying) and led to lower convergence. The lower interrater reliability for the traits of aesthetics and achievement-striving facets may be due to the fact that content related to these traits, particularly aesthetics, may not be typically gleaned from initial clinical interviews, making them more difficult to rate.
Intercorrelations
The FFMSS is also novel in that it allows for an assessment of maladaptive functioning at both poles of each domain. We found no evidence that the overall structure of the measure was substantially altered by the current methodology as the intercorrelations between the domains were relatively consistent with NEO PI-R intercorrelations from the normative data (Costa & McCrae, 1992), which suggests that the inclusion of maladaptive poles did not alter the manner in which the FFM domains relate to one another in a fundamental manner. In fact, the two correlations that differed
Few et al. significantly from the pattern reported in Costa and McCraes (1992) normative data (i.e., Conscientiousness and Neuroticism; Conscientiousness and Extraversion) differed such that the FFMSS correlations were smaller than that found for the FFM suggesting that the FFMSS results in a more orthogonal pattern of interrelations.
449 correlates. As a result, Miller and Lynam generated a revised FFMDependent PD prototype on the basis of a meta-analysis; the revised FFM Dependent PD profile, which deemphasizes the role of high Agreeableness and emphasizes the role of low Conscientiousness, manifested a stronger correlation with the current FFMSSDependent PD profile (r = .62) compared with the original prototype (r = .52). The FFMSS also evinced a substantial correlation with ObsessiveCompulsive PD, which is atypical in the FFM-PD literature (e.g., Miller et al., 2004). This strong correlation is likely due to the inclusion of maladaptive extremes at the high end of Conscientiousness, which are thought to be central to this PD. Haigler and Widiger (2001) demonstrated that the failure of certain FFM measures (i.e., NEO PI-R) to capture certain PDs, such as Schizotypal, Dependent, and OCPD, is due to a lack of items that reference maladaptivity at the high ends of Openness, Dependent, and OCPD, respectively. These authors found more substantial correlations between these three PDs and three FFM domains after experimentally altering NEO PI-R items to describe problematically high levels of these traits. The FFMSS, which includes maladaptivity at both the low and high ends of each FFM domain, did a much better job at capturing OCPD because of the changes to Conscientiousness. The same pattern was not found, however, for Schizotypal and Dependent PD, which still manifested limited correlations with Openness and Agreeableness, despite the inclusion of maladaptive levels of these two FFM domains.
450 FFMSS in clinical settings as it provides meaningful and novel information that is pertinent to client functioning and that may be inadequately assessed by current DSM-IV diagnostic constructs.
Assessment 17(4) feasible as part of an overall assessment summary. Therefore, it may not require additional clinician investment to rate clients effectively using the FFMSS. Another potential limitation relates to the characteristics of the current sample. Because the recruiting procedures were initially aimed at patients with Borderline and Avoidant PDs and patients with the absence of a PD diagnosis, the sample was overly weighted toward these forms of personality pathology. Despite this, the FFMSS correlated with the DSM-IV PDs in a manner that is consistent with previous studies (e.g., Samuel & Widiger, 2008). Finally, outside of the SNAP, the constructs explored here were all scored by the same group of expert raters, which raises the possibility that some of the current effect sizes are inflated due to shared method variance. In conclusion, the current results provide further support for the reliability and validity of the use of single-item assessments of the 30 facets and five domains pertinent to the FFM. In the current study, we demonstrated that the FFMSS can be used to rate patients with reasonable reliability following only limited training and that the FFMSS provides personality data that is relatively convergent with other self-report personality instruments, expert ratings of DSM-IV PDs from an FFM perspective, and clinician ratings of DSM-IV PDs. The ratings also demonstrated clinical utility in predicting concurrent impairment across a number of domains (e.g., romance, work). Future studies should examine further the reliability, validity, and utility of this measure and compare it to a similar rating form (FFMRF) to see which measure is superior for use in clinical and research settings.
Appendix
Five-Factor Model Score Sheet (Widiger & Spitzer, 2002) 7 = Problematic very high on the trait 6 = Problematic high on the trait (clear presence of clinically significant impairments) 5 = High on the trait (higher than the average, typical person; may or may not have minor impairments) 4 = Neither high nor low on the trait 3 = Low on the trait (lower than the average, typical person; may or may not have minor impairments) 2 = Problematic low on the trait (clear presence of clinically significant impairments) 1 = Problematic very low on the trait ? = Unable to estimate Circle number that applies or ?:
Neuroticism ? ? ? Anxiousness (fearful, apprehensive) Angry hostility: (angry, bitter) Depressiveness: (pessimistic, glum) 7 7 7 6 6 6 5 5 5 4 4 4 3 3 3 2 2 2 1 1 1 (relaxed, unconcerned, cool) (even-tempered) (optimistic) (continued)
Few et al.
451
Appendix (continued)
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? Self-consciousness: (timid, embarrassed) Impulsivity: (tempted, urgency) Vulnerability: (helpless, difficulty dealing with stress) Warmth: (affectionate, attached) Gregariousness: (sociable, outgoing) Assertiveness: (dominant, forceful) Activity: (vigorous, energetic, active) Excitement-seeking: (reckless, daring) 7 7 7 6 6 6 5 5 5 4 4 4 3 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 (self-assured, glib, shameless) (controlled, restrained) (stalwart, brave, fearless, unflappable) (Cold, aloof, indifferent) (Withdrawn, isolated) (Unassuming, quiet, resigned) (Passive, lethargic) (Cautious, monotonous, dull) (Placid, anhedonic) (practical, concrete) (unaesthetic, uninvolved) (constricted, constricted, alexythymic) (routine, habitual, stubborn) (pragmatic, rigid) (traditional, inflexible, dogmatic) (skeptical, cynical, suspicious, paranoid) (cunning, manipulative, deceptive) (stingy, selfish, greedy, exploitative) (oppositional, combative, aggressive) (confident, boastful, arrogant) (tough, callous, ruthless) (lax, negligent) (haphazard, disorganized, sloppy) (casual, undependable, unethical) (aimless, desultory) (hedonistic, negligent) (hasty, careless, rash)
Positive emotions: (high-spirited) 7 6 5 4 3 Openness Versus Closedness to Experience Fantasy: (dreamer, unrealistic, 7 6 5 4 3 imaginative) Aesthetic: (preoccupied, aberrant, 7 6 5 4 3 aesthetic) Feelings: (sensitive, responsive) 7 6 5 4 3 Actions: (unpredictable, unconventional) Ideas: (strange, odd, peculiar, creative) Values: (permissive, broad-minded) Trust: (gullible, trusting) Straightforwardness: (naive, honest) Altruism: (sacrificial, giving) Compliance: (docile, cooperative) Modesty: (meek, self-effacing, humble) Tender-mindedness: (soft, empathic) Competence: (perfectionistic, efficient) Order: (ordered, methodical, organized) Dutifulness: (rigid, reliable, dependable) Achievement: (workaholic, ambitious) Self-Discipline: (dogged, devoted) Deliberation: (ruminative, reflective) 7 7 7 6 6 6 5 5 5 4 4 4 3 3 3
7 6 5 4 Conscientiousness 7 6 5 4 7 6 5 4 7 7 7 7 6 6 6 6 5 5 5 5 4 4 4 4
Provide average of facet scores within each of the five domains to obtain global five-factor model description:
Neuroticism Extraversion Description for Five Broad Domains of Personality 7 6 5 4 3 2 7 6 5 4 3 2 1 1 Low neuroticism Introversion (continued)
452
Assessment 17(4)
Appendix (continued)
Openness Agreeableness Conscientiousness 7 7 7 6 6 6 5 5 5 4 4 4 3 3 3 2 2 2 1 1 1 Closedness Antagonism Low conscientiousness
Bagby, R. M., Rector, N. A., Bindseil, K., Dickens, S. E., Levitan, R. D., & Kennedy, S. H. (1998). Self-report ratings and informants ratings of personalities of depressed outpatients. American Journal of Psychiatry, 155, 437-438. Carroll, B. J., Fielding, J. M., & Blashki, T. G. (1973). Depression rating scales: A critical review. Archives of General Psychiatry, 28, 361-366. Clark, L. A. (1993). Manual for the Schedule for Nonadaptive and Adaptive Personality (SNAP). Minneapolis: University of Minnesota Press. Clark, L. A. (2007). Assessment and diagnosis of personality disorder: Perennial issues and an emerging reconceptualization. Annual Review of Psychology, 58, 227-257. Clark, L. A., Vorhies, L., & McEwen, J. L. (2002). Personality disorder symptomatology from the five-factor model perspective. In P. T. Costa & T. A. Widiger (Eds.), Personality disorders and the five-factor model of personality (2nd ed., pp. 125-148). Washington, DC: American Psychological Association. Cohen, J., & Cohen, P. (1983). Applied multiple regression/ correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Costa, P. T., & McCrae, R. R. (1992). Revised NEO personality inventory (NEO-PI-R) and NEO five-factor inventory (NEOFFI) professional manual. Odessa, FL: Psychological Assessment Resources. Costa, P. T., & Widiger, T. A. (Eds.). (2002). Personality disorders and the five-factor model of personality (2nd ed.). Washington, DC: American Psychological Association. Digman, J. M. (1990). Personality structure: Emergence of the five-factor model. Annual Review of Psychology, 41, 417-470. First, M. B., Gibbon, M., Spitzer, R. L., & Williams, J. B. W. (1997). Structured clinical interview for DSM-IV Axis-II personality disorders. Washington DC: American Psychiatric Press. Gosling, S. D., Rentfrow, P. J., & Swann, W. B., Jr. (2003). A very brief measure of the Big-Five personality domains. Journal of Research in Personality, 37, 504-528. Haigler, E. D., & Widiger, T. A. (2001). Experimental manipulation of NEO PI-R items. Journal of Personality Assessment, 77, 339-358. Hamilton, M. (1959). The assessment of anxiety states by rating. British Journal of Medical Psychology, 32, 50-55. Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery, and Psychiatry, 23, 56-62. Hopwood, C. J., Morey, L. C., Ansell, E. B., Grilo, C. M., Sanislow, C. A., McGlashan, T. H., . . . Skodol, A. E. (2009). The convergent and discriminant validity of the five-factor traits: Current and prospective social, work, and
Funding
The authors disclosed receipt of the following financial support for the research and/or authorship of this article: National Institute of Mental Health (NIMH) Grant R01 MH056888, Interpersonal Functioning in Borderline Personality to P. A. Pilkonis.
Notes
1. Several abbreviated measures of the five-factor model (FFM) have been developed. These include the Big Five Inventory (BFI; John & Srivastava, 1999), the NEO-Five Factor Inventory (NEO-FFI; Costa & McCrae, 1992), Sauciers (1994) Mini Markers, and the 10-Item Personality Inventory (TIPI; Gosling, Rentfrow, & Swann, 2003). One limitation of these measures, however, is that they do not assess the facets of the FFM (Mullins-Sweatt, Jamerson, Samuel, Olson, & Widiger, 2006). This is problematic in that the FFM facets help differentiate between both Axis I and Axis II disorders and are better able to capture the different personality disorders (PDs; see Lynam & Widiger, 2001). 2. The Five-Factor Model Score Sheet ([FFMSS] vs. the FFM rating Form [FFMRF]) was the only measure available at the time that this study was initiated. 3. The team of interviewers and expert judges comprised licensed psychologists, senior doctoral students (i.e., completing internship training), and masters-level clinicians (this team was composed of both male and female raters). Expert raters, outside of the primary interviewer, did not typically have any direct contact with the participants. Outside of the second author and fifth authors (both of whom provided a very small percentage of the primary ratings used in most of the current analyses), the interviewers and judges tended to have limited familiarity with the FFM. 4. There was some evidence of multicollinearity in these hierarchical models; however, we were less concerned with the significance of the individual betas than with the overall change in variance explained by the set of FFM domains. 5. We also tested whether the relations between the FFM domains and the impairment outcomes were curvilinear; there were no instances of statistically significant curvilinearity.
References
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.
Few et al.
recreational dysfunction. Journal of Personality Disorders, 23, 466-476. John, O. P., & Srivastava, S. (1999). The big five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed., pp. 102-138). New York, NY: Guilford Press. Klein, D. (2003). Patients versus informants reports of personality disorders in predicting 7-year outcome in outpatients with depressive disorders. Psychological Assessment, 15, 216-222. Lowe, J. R., & Widiger, T. A. (2009). Clinicians judgments of clinical utility: A comparison of the DSM-IV with dimensional models of general personality. Journal of Personality Disorders, 23, 211-229. Lynam, D. R., & Widiger, T. A. (2001). Using the five-factor model to represent the DSM-IV personality disorders: An expert consensus approach. Journal of Abnormal Psychology, 110, 401-412. Maier, W., Buller, R., Philipp, M., & Heuser, I. (1988). The Hamilton Anxiety Scale: Reliability, validity, and sensitivity to change in anxiety and depressive disorders. Journal of Affective Disorders, 14, 61-68. McCrae, R. R. (2008). A note on some measures of profile agreement. Journal of Personality Assessment, 90, 105-109. Miller, J. D., & Lynam, D. R. (2001). Structural models of personality and their relation to antisocial behavior: A meta-analytic review. Criminology, 39, 765-798. Miller, J. D., Pilkonis, P. A., & Clifton, A. (2005). Self- and otherreports of traits from the five-factor model: Relations to personality disorder. Journal of Personality Disorders, 19, 400-419. Miller, J. D., Pilkonis, P. A., & Mulvey, E. P. (2006). Treatment utilization and satisfaction: Examining the contributions of Axis II psychopathology and the five-factor model of personality. Journal of Personality Disorders, 4, 369-387. Miller, J. D., Reynolds, S. K., & Pilkonis, P. A. (2004). The validity of the five-factor model prototypes for personality disorders in two clinical samples. Psychological Assessment, 16, 310-322. Morse, J. Q., Hill, J., Pilkonis, P. A., Yaggi, K. E., Broyden, N., Stepp, S. D., . . . & Feske, U. (2009). Anger, preoccupied attachment, and domain disorganization in borderline personality disorder. Journal of Personality Disorders, 23, 240-257. Mullins-Sweatt, S. N., Jamerson, J. E., Samuel, D. B., Olson, D. R., & Widiger, T. A. (2006). Psychometric properties of an abbreviated instrument of the five-factor model. Assessment, 13, 119-137.
453
Oltmanns, T. F., Friedman, J. N. W., Fiedler, E. R., & Turkheimer, E. (2004). Perceptions of personality disorders based on thin slices of behavior. Journal of Research in Personality, 38, 216229. Reynolds, S. K., & Clark, L. A. (2001). Predicting dimensions of personality disorder from domains and facets of the five-factor model. Journal of Personality, 69, 199-222. Samuel, D. B., & Widiger, T. A. (2004). Clinicians personality descriptions of prototypic personality disorders. Journal of Personality Disorders, 18, 286-308. Samuel, D. B., & Widiger, T. A. (2006). Clinicians judgments of clinical utility: A comparison of the DSM-IV and five-factor models. Journal of Abnormal Psychology, 115, 298-308. Samuel, D. B., & Widiger, T. A. (2008). A meta-analytic review of the relationships between the five-factor model and DSMIV-TR personality disorders: A facet level analysis. Clinical Psychology Review, 28, 1326-1342. Saucier, G. (1994). Mini-markers: A brief version of Goldbergs unipolar big five markers. Journal of Personality Assessment, 63, 506-516. Spitzer, R. L. (1983). Psychiatric diagnosis: Are clinicians still necessary? Comprehensive Psychiatry, 24, 399-411. Sprock, J. (2002). A comparative study of the dimensions and facets of the five-factor model in the diagnosis of cases of personality disorder. Journal of Personality Disorders, 16, 402-423. Trull, T. J., & Widiger, T. A. (1997). Structured interview for the five-factor model of personality (SIFFM): Professional manual. Odessa, FL: Psychological Assessment Resources. Trull, T. J., Widiger, T. A., Useda, J. D., Holcomb, H., Doan, B., Axelrod, S. R., . . . Gershuny, B. S. (1998). A structured interview for the assessment of the five-factor model of personality. Psychological Assessment, 10, 229-240. Watson, D., Clark, L. A., & Chmielewski, M. (2008). Structures of personality and their relevance to psychopathology. II. Further articulation of a comprehensive unified trait structure. Journal of Personality, 76, 1545-1586. Whiteside, S. P., & Lynam, D. R. (2001). The five factor model and impulsivity: Using a structural model of personality to understand impulsivity. Personality and Individual Differences, 30, 669-689. Widiger, T. A. (2004). Five-Factor Model Rating Form. Unpublished measure. Widiger, T. A., & Spitzer, R. L. (2002). Five-Factor Model Score Sheet. Unpublished measure.